U.S. patent application number 09/896228 was filed with the patent office on 2003-03-13 for system and method for router virtual networking.
Invention is credited to Visser, Lance Arnold, Watson, Thomas Lee.
Application Number | 20030051048 09/896228 |
Document ID | / |
Family ID | 25405843 |
Filed Date | 2003-03-13 |
United States Patent
Application |
20030051048 |
Kind Code |
A1 |
Watson, Thomas Lee ; et
al. |
March 13, 2003 |
System and method for router virtual networking
Abstract
A host router is logically partitioned into virtual router
domains that manage independent processes and routing application
copies but share a common operating system. Each v-net manages an
independent set of sockets and host router interfaces, each
associated with only one v-net at one time, but interchangeably
repartitionable Traffic is removed from an interface during
repartitioning. Duplicate arrays of global variables copied to each
v-net are accessed by macro references. A v-net facility can
separate route tables used internally from the externally visible
route tables and can avoid conflicts between internal and external
IP addresses that share the same identifier. For example a common
FreeBSD operating system supports a dynamic routing protocol (DRP)
application. Each v-net runs an independent copy of the DRP
software and is logically independent. A failure in one DRP copy
does not adversely affect other copies.
Inventors: |
Watson, Thomas Lee;
(Richardson, TX) ; Visser, Lance Arnold; (Dallas,
TX) |
Correspondence
Address: |
FULBRIGHT & JAWORSKI L.L.P.
William B. Tiffany
Suite 2800
2200 Ross Avenue
Dallas
TX
75201
US
|
Family ID: |
25405843 |
Appl. No.: |
09/896228 |
Filed: |
June 29, 2001 |
Current U.S.
Class: |
709/238 |
Current CPC
Class: |
H04L 45/586 20130101;
H04L 45/00 20130101 |
Class at
Publication: |
709/238 |
International
Class: |
G06F 009/46; G06F
015/173 |
Claims
What is claimed is:
1. A system of virtual router domains comprising: a host router
running a common operating system; a plurality of virtual router
domains and processes logically partitioned within said host
router, each said virtual router domain having a unique domain ID
address and an independent replica array of all virtualized
variables across said common operating system, each said process
running in a said virtual router domain independently of all other
said virtual router domains on top of said common operating system;
and said global variables being accessed by macro references in
each said virtual router domain.
2. The system of claim 1 wherein said common operating system runs
on a master control processor within said host router.
3. The system of claim 2 wherein said common operating system is a
version of FreeBSD.
4. The system of claim 1 wherein said common operating system
manages the reporting of hardware failures across all virtual
router domains of said host router.
5. The system of claim 1 wherein said plurality of processes
comprise routing software applications.
6. The system of claim 5 wherein said plurality of processes
comprise independent plural identical copies of at least one said
process.
7. The system of claim 5 wherein said plurality of processes
comprise a copy of a dynamic routing protocol (DRP) software
application.
8. The system of claim 1 wherein said plurality of processes
comprise a copy of a SNMP application.
9. The system of claim 1 wherein said macros generate an array of
said global variables when said virtual router domain is configured
in.
10. The system of claim 9 wherein said macros generate scalar
global variables when said virtual router domain is
deconfigured.
11. The system of claim 1 further comprising a plurality of
interfaces partitioned interchangeably among said virtual router
domains, such that a particular interface is associated with only
one such virtual router domain at one time, but can be
repartitioned among said virtual router domains to reconfigure said
host router.
12. The system of claim 11 wherein during said reconfiguring
network traffic is removed from said interfaces that are
repartitioned.
13. The system of claim 11 wherein said interface contains the
unique domain ID address of said virtual router domain with which
said interface is associated.
14. The system of claim 13 wherein said interface is an interface
port of said host router.
15. The system of claim 14 wherein said host router comprises at
least 320 said interface ports.
16. The system of claim 14 further comprising a socket created by
at least one said process, said socket being associated exclusively
with the virtual router domain in which it is created and
containing said unique domain ID address of said domain in which it
is created.
17. The system of claim 16 wherein multiple sockets are created by
said at least one process in at least one said virtual router
domain, such that each of said multiple sockets is associated
exclusively with the virtual router domain in which said socket is
created.
18. The system of claim 17 wherein said at least one process is
movable from one said virtual router domain to a different said
virtual router domain, such that said at least one process creates
a said socket in each of at least two said virtual router
domains.
19. The system of claim 17 wherein a particular socket associated
with a particular virtual router domain is applied exclusively to
live traffic networking independently of any other said virtual
router domain of said host router.
20. The system of claim 17 wherein a particular socket associated
with a particular virtual router domain is applied exclusively to a
test bed operation independently of any other said virtual router
domain of said host router.
21. The system of claim 16 wherein each of said virtual router
domains maintains an independent routing table.
22. The system of claim 21 wherein each said socket uses the
routing table of said virtual router domain in which said socket is
created.
23. The system of claim 21 wherein said two distinct virtual router
domains use the same Internet Protocol (IP) addresses without
conflicting.
24. The system of claim 21 wherein one particular virtual router
domain within said host router contains routing tables exclusively
for internal interface addresses within said host router
independently of any other said virtual router domain of said host
router.
25. The system of claim 21 wherein a particular virtual router
domain within said host router contains routing tables exclusively
for interfaces externally visible from outside said host router
independently of any other said virtual router domain of said host
router.
26. The system of claim 21 wherein a failure of one of said
plurality of said virtual router domains does not adversely affect
a different one of said plurality of said virtual router
domains.
27. A method of logically partitioning a host router into virtual
router domains, comprising configuring the kernel of a single
common operating system running in said host router; configuring in
a plurality of virtual router domains within said host router;
identifying each said virtual router domain by a unique domain
index number; generating an independent identical set of replica
arrays of global variables for each virtual router domain; and
associating a process with each said virtual router domain of said
host router, such that said processes run in said virtual router
domains independently of one another on top of said single common
operating system of said host router.
28. The method of claim 27 wherein said global variables are
generated by macros.
29. The method of claim 28 wherein said macros generate arrays of
global variables when said virtual router domain is configured in
within said host router.
30. The method of claim 29 wherein said macros generate scalar
global variables when said virtual router domain is
deconfigured.
31. The method of claim 27 wherein said single common operating
system runs on a master control processor within said host
router.
32. The method of claim 31 wherein said single common operating
system is a version of FreeBSD.
33. The method of claim 27 wherein said processes comprise routing
software applications.
34. The method of claim 33 further comprising independently running
plural identical copies of at least one said process.
35. The method of claim 33 wherein said processes comprise a copy
of a dynamic routing protocol (DRP) software application.
36. The method of claim 27 wherein each said process manages an
instantiation of a common networking code.
37. The method of claim 27 further comprising partitioning a
plurality of host router interfaces interchangeably among said
virtual router domains, such that a particular interface is
associated with only one such virtual router domain at one
time.
38. The method of claim 37 further comprising repartitioning said
plurality of interfaces among said virtual router domains, such
that said host router is reconfigured.
39. The method of claim 38 wherein during said reconfiguring
network traffic is removed from said interfaces that are
repartitioned.
40. The method of claim 37 wherein said interface contains the
unique domain index number of said virtual router domain with which
said interface is associated.
41. The method of claim 40 wherein said interface is an interface
port of said host router.
42. The method of claim 41 wherein said host router comprises at
least 320 said interface ports.
43. The method of claim 41 wherein said process creates a socket,
such that said socket is associated permanently and exclusively
with the virtual router domain in which it is created and
containing said unique domain index number of said domain in which
it is created.
44. The method of claim 43 wherein multiple sockets are created by
at least one said process in at least one said virtual router
domain, such that each of said multiple sockets is associated
permanently and exclusively with the virtual router domain in which
said socket is created.
45. The method of claim 44 further comprising moving said at least
one process from one said virtual router domain to a different said
virtual router domain, such that said at least one process creates
a said socket in each of at least two said virtual router
domains.
46. The method of claim 44 wherein said process maintains a file
descriptor table containing pointers to said sockets associated
with said virtual router domain.
47. The method of claim 46 wherein a particular socket associated
with a particular virtual router domain is applied exclusively to
live traffic networking independently of any other said virtual
router domain of said host router.
48. The method of claim 46 wherein a particular socket associated
with a particular virtual router domain is applied exclusively to a
test bed operation independently of any other said virtual router
domain of said host router.
49. The method of claim 43 wherein each of said virtual router
domains maintains an independent routing table.
50. The method of claim 49 wherein each said socket uses the
routing table of said virtual router domain in which said socket is
created.
51. The method of claim 49 wherein said two distinct virtual router
domains use the same Internet Protocol (IP) addresses without
conflicting.
52. The method of claim 49 wherein one particular virtual router
domain within said host router contains routing tables exclusively
for internal interface addresses within said host router
independently of any other said virtual router domain of said host
router.
53. The method of claim 49 wherein a particular virtual router
domain within said host router contains routing tables exclusively
for interfaces externally visible from outside said host router
independently of any other said virtual router domain of said host
router.
54. The method of claim 49 wherein a failure of one of said
plurality of said virtual router domains does not adversely affect
a different one of said plurality of said virtual router domains.
Description
BACKGROUND
[0001] Transmission Control Protocol (TCP) is an underlying
connection protocol that is typically used for all types of network
communication. A route is essentially the mapping of an IP address
to an egress port of a router. Different network routers set up
connections with their peer routers using operating systems, for
example Border Gateway Protocol (BGP) over TCP or OSPF (Open
Shortest Path First) over Internet Protocol (IP) to determine that
they get route information from their peers, allowing them to
construct essentially an internal map of the network and to select
the route that they should use, as well as verification that their
peers are operating correctly. This is accomplished by sending
various keep-alive packets back and forth to make sure that their
peers are still correctly functioning. Routes are used internally
within a router, for example a Master Control Processor (MCP)
communicates through an Ethernet control network (CNET) within a
router with the shelf control processors, each of which have
individual IP addresses. Processes including routing applications,
for example Dynamic Routing Protocol (DRP), run on these operating
systems. Sockets are end points of communication associated with a
process. A particular process can have more than one socket.
[0002] In a router with a large number of ports, for example 320
ports, that communicates with peer routers, it is advantageous to
subdivide that single large router logically into several smaller
virtual routers, each of which can be individually configured.
There can be separate departments in a large company, or an
Internet provider wanting to partition a large router among
clients, for example for security reasons. However, previous
implementations of subdividing routers having large numbers of
ports have been cumbersome.
SUMMARY OF THE INVENTION
[0003] The present invention is directed to a system and method
which logically partition a host router into virtual router domains
that run independent processes and routing application copies but
share a common operating system. Each v-net domain manages an
independent set of interface ports. Each process manages an
independent set of sockets.
[0004] In some embodiments a v-net domain architecture is used to
partition a host router. Some v-net domains support virtual
routers, whereas other v-net domains support only internal router
processes and management applications. Thus, not every v-net domain
supports a virtual router. A single v-net domain can support more
than one process. A v-net facility can advantageously separate
route tables used internally from the externally visible routes,
making network management easier and more transparent. With
separate v-net domains for example, the IP address of an internal
shelf control processor does not conflict with the same IP address
that is assigned elsewhere on the Internet. In a v-net
implementation, duplicate arrays of global variables are
instantiated in each virtual router domain and are accessed by
macro references.
[0005] A common FreeBSD operating system running on the MCP
supports a dynamic routing protocol (DRP) application. Each new
virtual router is independently managed by its own copy of the DRP
application for as many virtual routers as exist. If something goes
awry in one DRP copy, it does not affect other copies. Each v-net
domain manages a separate set of the interfaces associated with the
host router, which provide connections to peer routers. For
example, if a host router has 320 ports, one v-net domain can
manage 120 ports or interfaces, and another v-net domain can manage
another 120 ports. All of these ports and interfaces can be
interchangeably partitioned. For each Synchronous Optical Network
(SONET) port on a line card, there is an interface (IF) data
structure in FreeBSD that represents that SONET port. Any interface
can be associated with only one v-net at one time, but can be moved
among v-nets to reconfigure the host router. Traffic is removed
from an interface while it is being moved. At a high level the host
router is partitioned, and each partition normally is managed by an
independent copy of the DRP software. In an administrative sense,
each of these partitions is logically independent.
[0006] Certain activities are still managed across the entire host
router, for example failure reporting of hardware in the host
router, which is machine specific, and therefore is a resource
shared by all of the partitions.
[0007] This partitioning also allows the routes between the
individual components such as the line cards and processors
internal to a router to be contained in route tables separate from
externally visible routes. Partitioning the router also facilitates
testing, such that one partition might be used for normal network
traffic and another might be used to test for example new software
or new network configurations for new types of protocols.
Additionally, a degree of redundancy is achieved, such that failure
of one partition generally does not adversely affect another
partition sharing the same host router.
[0008] Various aspects of the invention are described in co-pending
and commonly assigned U.S. application Ser. No. 09/703,057,
entitled "System And Method For IP Router With an Optical Core,"
filed Oct. 31, 2000, the disclosure of which has been incorporated
herein by reference.
[0009] The foregoing has outlined rather broadly the features and
technical advantages of the present invention in order that the
detailed description of the invention that follows may be better
understood. Additional features and advantages of the invention
will be described hereinafter which form the subject of the claims
of the invention. It should be appreciated by those skilled in the
art that the conception and specific embodiment disclosed may be
readily utilized as a basis for modifying or designing other
structures for carrying out the same purposes of the present
invention. It should also be realized by those skilled in the art
that such equivalent constructions do not depart from the spirit
and scope of the invention as set forth in the appended claims. The
novel features which are believed to be characteristic of the
invention, both as to its organization and method of operation,
together with further objects and advantages will be better
understood from the following description when considered in
connection with the accompanying figures. It is to be expressly
understood, however, that each of the figures is provided for the
purpose of illustration and description only and is not intended as
a definition of the limits of the present invention.
BRIEF DESCRIPTION OF THE DRAWING
[0010] For a more complete understanding of the present invention,
reference is now made to the following descriptions taken in
conjunction with the accompanying drawing, in which:
[0011] FIG. 1 is a logical diagram illustrating the principles of
router virtual networking, according to an embodiment of the
present invention.
DETAILED DESCRIPTION
[0012] In embodiments of the present invention, a host network
router is logically partitioned into multiple virtual networking
domains sharing a common operating system. FIG. 1 is a logical
diagram illustrating the principles of router virtual networking,
according to an embodiment of the present invention. In the
implementation of FIG. 1, a host router 10 is logically partitioned
into v-net domains 12, 14, and 16 that are associated with
networking systems. Each v-net 12, 14, 16 has a unique v-net ID
address 13, 15, 17, in accordance with network protocols. Host
router 10 and each of v-nets 12, 14, 16 are further logically
subdivided into two spaces, shown in FIG. 1 separated horizontally
by a solid line, namely a user level 18 and a kernel level 20 of
the shared common operating system (OS), for example a version of
FreeBSD. The present FreeBSD operating system runs on the host
router Master Control Processor (MCP), described for example in
U.S. application Ser. No. 09/703,057, entitled "System And Method
For IP Router With an Optical Core," filed Oct. 31, 2000, cited
above, the disclosure of which has been incorporated herein by
reference, and the dynamic routing protocol (DRP) application
software runs on top of FreeBSD.
[0013] An operating system contains within it logical notions
called processes 22-26, for example Internet Management Application
22, DRP 23, 25, or Simple Network Management Protocol (SNMP) agent
application 24, 26, running on v-nets 12, 14, and 16. Different
individual v-nets can manage the same, different, single, or
multiple processes. V-net domains 14 and 16, each running DRP and
SNMP processes, are virtual routers, whereas v-net domain 12,
running only an internal management application, is not a virtual
router. The present FreeBSD operating system supports multiple
processes, among which are DRP 23, 25, SNMP 24, 26, and Internal
Management Application 22. Each process occupies some user level
space 18 and also some operating system kernel level space 20. User
level space 18 includes the application and the values of all the
application variables (not shown in FIG. 1), whereas OS or kernel
level space 20 of the process includes internal data that the
kernel maintains with each process. Typical examples of internal
kernel data include descriptors or descriptions of open files and
the ID of the user that owns the process, attributes that are added
to each process associated with a particular v-net.
[0014] Among other things associated with a particular v-net are
interfaces, for example interfaces 42-1 through 42-3 associated
with v-net 12. An interface represents for example a particular
physical hardware Ethernet card, gigabit Ethernet card, or SONET
line card interconnected with a remote router. This allows
partitioning of host router interfaces, such that for example
interfaces 42-1 through 42-3 contain v-net ID 13 of v-net 12 with
which they are associated. V-net domain 12 maintains an interface
list 42-0 pointing to interfaces 42-1 through 42-3. Similarly v-net
domain 14 maintains an interface list 43-0 pointing to interfaces
43-1 through 43-3 carrying v-net ID 15 of v-net domain 14, and
v-net domain 16 maintains an interface list 45-0 pointing to
interfaces 45-1 through 45-3 carrying v-net ID 17 of v-net domain
16.
[0015] Each process 22-26 can create sockets, which are end points
of communication associated with a process, for example sockets
32-1 through 32-3 associated with process 22 in v-net domain 12. A
particular process can have more than one socket. Each socket has a
v-net ID associated with it, for example sockets 32-1 through 32-3
each contain v-net ID 13 of v-net 12. In v-net 12, management
application 22 maintains a descriptor table, for example file
descriptor table 32-0 of v-net 12, holding references to sockets
32-1 through 32-3 and to files, which are each associated with
specific application 22. Similarly, in v-net 14, DRP application 23
maintains descriptor table 33-0, holding references to sockets 33-1
through 33-3 and to files associated with application 23, and SNMP
application 24 maintains descriptor table 34-0 holding references
to sockets 34-1 through 34-3 and to files associated with
application 24. Likewise in v-net 16, DRP application 25 maintains
descriptor table 35-0, holding references to sockets 35-1 through
35-3 and to files associated with application 25, and SNMP
application 26 maintains descriptor table 36-0 holding references
to sockets 36-1 through 36-3 and to files associated with
application 26.
[0016] Sockets are partitioned basically according to the domain in
which communication takes place. Each of the things done to the
socket is interpreted in the context of the particular v-net in
which the socket is created, and therefore the socket carries that
particular v-net identifier. The process has a v-net identifier,
because when a process creates a new socket, which it is able to
do, each socket that it creates is then created in a process of
that v-net identifier. For example, if a process is associated with
v-net 0 creates a socket, then that socket is automatically
associated with v-net 0, gets its routing tables from v-net 0, and
can then use all of the interfaces that are assigned to v-net 0. A
process can, however, change its v-net identifier and thereby its
v-net association, for example by moving logically from v-net 0 to
v-net 1, and can then create a new socket associated with v-net 1,
which uses routing tables and interfaces of v-net 1, which are
disjoint with the interfaces for v-net 0.
[0017] Once a socket is created, it cannot be moved to another
v-net, but remains in the domain in which it was created. However,
a process, by changing its v-net identifier, can then create
sockets in multiple domains. Consequently, a process can
essentially communicate across domains by creating a socket in each
one, but each socket, throughout its existence, is fixed in its
original domain. Multiple sockets created by a process are
distinctly different from a single socket that is simply
interpreted in different ways. For example a single process can
create ten distinct sockets in one domain and five distinct sockets
in another domain. For example, socket 35-4 is created in v-net
domain 12 by DRP application 25 and carries v-net ID 13, although
socket 35-4 is referenced in descriptor list 35-0 of DRP
application 25, which is now in v-net domain 16. Likewise, socket
33-4 is created in v-net domain 12 by DRP application 23 and thus
carries v-net ID 13, although socket 33-4 is referenced in
descriptor list 33-0, which is now in v-net domain 14. A socket is
destroyed when a process exits or when a process closes down the
communication end point represented by that socket. After a socket
is destroyed, it is no longer associated with any domain, and the
memory associated with it is freed.
[0018] If for example v-net 14 and v-net 16 are two networking
domains of host router 10, and if v-net 14 is a production network
carrying live traffic with production code in it, or production
network connections carrying real customer traffic, then a socket
associated with v-net 14 is operating in that v-net's space and has
routing tables 48 for that v-net to route live traffic.
Consequently, if the socket were to select a particular IP address,
that IP address would use production routing tables 48. A different
socket in a different v-net 16 is for example used for a small test
bed and contains a different set of routing tables 50. Accordingly,
when a message is sent on v-net 16 with an IP address, that IP
address is interpreted in the context of v-net 16 running the small
test bed.
[0019] Global variables are variables that are accessible to all
the various logical contexts or threads of execution that are
running concurrently within an operating system. Thus a global
variable is not on the stack of a particular thread. Accordingly,
all global variables are available to every process that is running
within the operating system. Global variables include at least at
the top level, for example, the IP address of a machine or a copy
of the routing tables so that a process knows where to send
packets. There are a certain set of global variables associated
with the networking code, and in order to make the networking codes
support partitioning, the set of global variables associated with
networking are replicated, one copy 47 for each v-net domain, such
that the operating system effectively contains, rather than one
copy of the networking data structures, N instantionations of the
networking stack, replicating all the various functions of the
networking code, including replicated routing tables and replicated
TCP control blocks linked together throughout the basic data
structure. Thus, effectively all of the important variables in the
networking system are replicated, so that they can be independently
managed. This can be thought of as an operating system with N
instantiations of the networking system.
[0020] The basic approach of the v-net code is to take global
variables that need to be replicated for each v-net domain, and to
make an array of them. As an example tcpstat, the tcp statistics
structure, is declared in tcp_var.h struct tcpstat { . . . } and
defined in tcp_input. c as struct tcpstat tcpstat. To have a
separate set of statistics for each v-net domain requires changing
the definition to struct tcpstat tcpstat[NVNET] and changing all
references to index by the appropriate v-net domain number.
[0021] To make v-net facility a configuration option, the
declarations and references are encapsulated in macros. The macros
generate arrays when v-nets are configured in and scalars when
v-nets are deconfigured. As an example the tcpstat declaration
becomes VDECL (struct tcpstat, tcpstaT), in which the first macro
argument is the type, and the second macro argument is the name. It
will be noted that the variable name is changed from tcpstat to
tcpstaT. This convention is followed throughout the global variable
generation, i.e., variables that are virtualized and global across
more than one file are changed to have the final letter in their
name capitalized. This is done for three reasons:
[0022] 1) to differentiate global variables from local variables
and/or types of the same name for readability,
[0023] 2) to ensure that all references to global variables are
fixed appropriately (by causing a compile error if the variable
name is not changed); and
[0024] 3) to denote global variables plainly for possible future
changes.
[0025] References to virtualized variables are made using one of
two macros,_v(name), or _V(name, index), where name is the variable
name and index is the v-net domain index to be used. The macro _v
uses a per CPU global index variable vnetindex. It will be noted
that all references to virtualized variables must be made with
these macros, without exception, so that the references are correct
without requiring #ifdef's when v-nets are configured or
deconfigured.
[0026] In addition to defining a methodology that handles
virtualization of variables, a selection is needed of the correct
set of global variables to be replicated for each v-net domain, and
the replicated variables need to be correctly referenced by macros
in the appropriate v-net domain. For example, global variables can
be identified by using a script that analyzes object (_o) files for
the global variables they define, by code inspection, or by
information from other sources (see for example the tables of
global variables in TCP/IP Illustrated, Volume 2: The
Implementation, Gary R. Wright and W. Richard Stevens,
Addison-Wesley 1995, p. 64, 97, 128, 158, 186, 207, 248, 277, 305,
340, 383, 398, 437, 476, 572, 680, 715, 756, 797, 1028, and
1051).
[0027] The following Appendix A is basically a table of the global
variables that are virtualized in some implementations, listing the
name and the purpose of the variable. The variables that are
virtualized are generally marked "virtualized" in the table.
Although virtualized variables shown in the table are usually
marked "virtualized," other variables in the table have been
analyzed but excluded from virtualization. All of the "virtualized"
variables are essentially replicated, such that each v-net
maintains its own set of these variables. Then macros, program
conventions that allow textural substitution, are provided, such
that everywhere a global variable is accessed, a replacement access
is a macro reference selected from the correct set of variables
based on the correct v-net.
[0028] In the present embodiment, multiple networking domains are
implemented by the same operating system, unlike previous
approaches, in which for example a computer is subdivided into
virtual domains that partition the hardware and run separate
operating systems in each domain.
appendix A. Variable Analysis
[0029]
1 NOTE: In the Analysis/Disposition column, "Virtualized" means the
variable becomes an array when vnets are configured (see the
description above); "Invariant" means a separate instance of the
variable is not needed for different vnet domains; and "Not
Virtualized" means there was a choice about virtualization (e.g.,
whether a Tunable could have a different value in different
domains), but the choice was made not to virtualize the variable.
Defining Analysis/ Variable Data Type File Description Disposition
Head static struct igmp.c Head of router_info linked list.
Virtualized. router_info* Addmask_key static char* radix.c
Temporary storage for Invariant. m_addmask. arp_allocated static
int if_ether.c Total number of llinfo_arp Virtualized. structures
allocated. arp_inuse static int if_ether.c Current number of
llinfo_arp Virtualized. structures in use. arp_maxtries static int
if_ether.c Tunable. Maximum number of Tunable. Not retries for an
arp request. virtualized. arp_proxyall static int if_ether.c
Tunable. Enables forming a Tunable. Not proxy for all arp requests.
virtualized. arpinit_done static int if_ether.c Indicates
initialization is done. Invariant. Initialization handles all
vnets. arpintrq struct ifqueue if_ether.c Arp interrupt request
queue. Invariant. Shared by all vnets. Vnet switching when pulled
off queue. arpt_down static int it_ether.c Tunable. No. of seconds
Tunable. Not between ARP flooding virtualized. algorithm. arpt_keep
static int it_ether.c Tunable. No. seconds ARP Tunable. Not entry
valid once resolved. virtualized. arpt_prune static int it_ether.c
Tunable. No. seconds between Tunable. Not checking ARP list.
virtualized. bpf_bufsize static int bpf.c Tunable. Tunable. Not
virtualized. bpf_cdevsw static struct bpf.c Table of entry point
function Invariant. cdevsw pointers. bpf_devsw_installed static int
bpf.c Initialization flag. Invariant. bpf_dtab static struct bpf.c
Descriptor structure, one per Invariant. bpf_d open bpf device.
(NBPFILTER) bpf_dtab_init static int bpf.c Another initialization
flag. Invariant. bpf_iflist static struct bpf.c Descriptor
associated with each Invariant. bpf_if attached hardware interface.
clns_recvspace static u_long raw_clns.c Constant (patchable).
Amount Not virtualized. of receive space to reserve in socket.
clns_sendspace static u_long raw_clns.c Constant (patchable).
Amount Not virtualized. of send space to reserve in socket.
clns_usrreqs struct pr_usrreqs raw_clns.c Function pointers for
clns user Invariant. requests. clnsg struct clnsglob raw_clns.c
Global state associated with Virtualized. ray_clns.c, including
list heads and counters. clnsintrq struct ifqueue raw_clns.c Clns
interrupt request queue. Invariant. Shared by all vnets. Vnet
switching done when removed from queue. clnssw struct protosw
raw_clns.c Pointers to protocol entry Invariant. points &
associated data. counter static u_int64_t ip_fw.c Counter for
ipfw_report. Virtualized. div_recvspace static u_long ip_divert.c
Amount of receive space to Invariant. reserve in socket.
div_sendspace static u_long ip_divert.c Amount of send space to
Invariant. reserve in socket divcb static struct ip_divert.c Head
of inpcb structures for Virtualized. inpcbhead divert processing.
divcbinfo static struct ip_divert, c Pcbinfo structure for divert
Virtualized. inpcbinfo processing dst static struct bpf.c Sockaddr
prototype. Invariant. sockaddr err_prefix char[] ip_fw.c Constant
string for printfs. Invariant. etherbroadcastaddr u_char [6]
if_ethersubr.c Constant. Ethernet broadcast Invariant. link
address. expire_upcalls_ch static struct ip_mroute.c Callout handle
for Virtualized. callout_handle expire_upcalls. fcstab static
u_short ppp_tty.c Constant. Table for FCS Invariant. [256] lookup.
frag_divert_port static u_short ip_input.c Divert protocol port. ?
Conditionally compiled iwith IPDIVERT. fw_debug static int ip_fw.c
Tunable. Enables debug print. Not virtualized. fw_one_pass static
int ip_fw.c Tunable. Enables accepting Not virtualized. packet if
passes first test. fw_verbose static int ip_fw.c Tunable; controls
verbosity of Not virtualized. firewall debugging messages.
fw_verbose_limit static int ip_fw.c Tunable. Limits amount of Not
virtualized. logging. have_encap_tunnel static int ip_mroute.c
Indicates presence of an Virtualized. encapsulation tunnel.
icmpbmcastecho static int ip_icmp.c Tunable flag. Disables Not
virtualized. broadcasting of ICMP echo and timestamp packets.
icmpdst static struct ip_icmp.c Saves the source address for
Virtualized. sockaddr_in ifaof_ifpforaddr. icmpgw static struct
ip_icmp.c Holds the ip source address in Virtualized. sockaddr_in
icmp_input. May not be necessary icmplim static int ip_icmp.c
Tunable. ICMP error-response Not virtualized. band with limiting
sysctl. icmpmaskrepl static int ip_icmp.c Tunable flag. Enables
ICMP Not virtualized. mask replacement. icmpprintfs int ip_icmp.c
Enables printfs in icmp code. Not virtualized. icmpsrc static
struct ip_icmp.c Holds the ip dest address in Virtualized.
sockaddr_in icmp_input. May not be necessary icmpstat static struct
ip_icmp.c Icmp statistics. Virtualized. icmpstat if_indeX int if.c
Number of configured Virtualized. interfaces. if_indexliM static
int if.c Number of entries in Virtualized. ifnet_addrS array. ifneT
struct ifnethead if.c Head of list of ifnet structures.
Virtualized. ifnet_addrS struct iffaddr** if.c Array of pointers to
link level Virtualized. interface addresses. ifqmaxlen int if.c
Constant. Maximum queue Invariant. length for interface queue.
igmp_all_hosts_group static u_long igmp.c Host order of Invariant.
INADDR_ALLHOSTS_GROUP constant igmp_all_rtrs_group static u_long
igmp.c Host order of Invariant. INADDR_ALLRTS_GROUP constant.
igmp_timers_are_running static int igmp.c Flag indicating any igmp
timer Virtualized. is active. igmprt static struct route igmp.c
Temporary variable. Invariant. igmpstat static struct igmp.c Igmp
statistics. Virtualized. igmpstat in_ifaddrheaD struct ip_input.c
Head of in_ifaddr structure list. Virtualized. in_ifaddrhead
in_interfaces static int in.c Incremented each time a non-
Invariant. loopback interface is added to Never read.
in_ifaddrheaD. Not read. Dead code. in_multiheaD struct in.c Head
of list of Virtualized. in_multihead in_multistructures (multicast
address). inetclerrmap u_char [] ip_input.c Array of constants
(error Invariant. numbers). inetdomain struct domain in_proto.c
Pointers to switch table, Invariant. initialization, etc. for
internet domain. inetsw struct protosw in_proto.c Pointers to entry
points for Invariant. various internet protocols. inited static int
if.c Flag indicating initialization Invariant. has been performed.
Initialization does all vnets. ip_acceptsourceroute static int
ip_input.c Tunable flag. Enables Tunable. Not acceptance of source
routed virtualized. packets. ip_defttl int ip_input.c Tunable.
Default time to live Tunable. Not from RFC 1340. virtualized.
ip_divert_cookiE u_intl6_t ip_divert.c Cookie passed to user
process. Virtualized. ip_divert_porT u_short ip_divert.c Global
"argument" to Virtualized. div_input. Used to avoid changing
prototype. ip_dosourceroute static int ip_input.c Tunable flag.
Enables acting as Tunable. Not a router. virtualized. ip_fw_chaiN
struct ip_fw_head ip_fw.c Head of ip firewall chains. Virtualized.
ip_fw_chk_ptr ip_fw_chk_t* ip_input.c IP firewall function callout
Invariant. pointer; value depends on loading fw module.
ip_fw_ctl_ptr ip_fw_ctl_t* ip_input.c IP firewall function callout
Invariant. pointer; value depends on loading fw module.
ip_fw_default_rulE struct ip_fw.c Pointer to default rule for
Virtualized. ip_fw_chain* firewall processing. ip_fw_fwd_addR
struct ip_input.c IP firewall address. Virtualized. sockaddr_in*
ip_ID u_short ip_output.c IP packet identifier Virtualized.
(increments). ip_mcast_src ulong (*)(int) ip_mroute.c Pointer to
function; selection Invariant. depends on compile options.
ip_mforward int(*)(struct ip*, ip_mroute.c Function pointer set by
module Invariant. struct ifnet*, . . . ) installation. ip_mrouteR
struct socket* ip_mroute.c Socket of multicast router Virtualized.
program. ip_mrouter_done int (*)(void) ip_mroute.c Function pointer
set by module Invariant. installation. ip_mrouter_get int
(*)(struct ip_mroute.c Function pointer selected by Invariant.
socket*, struct compile options. sockopt*) ip_mrouter_set int
(*)(struct ip_mroute.c Function pointer selected by Invariant.
socket*, struct compile options. sockopt*) ip_nat_clt_ptr
ip_nat_ctl_t* ip_input.c IP firewall function callout Invariant.
hook; set by module install. ip_nat_ptr ip_nat_t* ip_input.c IP
firewall function callout Invariant. hook; set by module install.
ip_nhops static int ip_input.c Hop count for previous source
Virtualized. route. ip_protox u_char ip_input.c Maps protocol
numbers to Invariant. [PROTO_MAX] inetsw array. ip_rsvpD struct
socket* ip_input.c Pointer to socket used by rsvp Virtualized.
daemon. ip_rsvp_on static int ip_input.c Boolean indicating rsvp is
Virtualized. active. ip_srcrt struct ip_srcrt ip_input.c Previous
source route. Virtualized. ipaddR struct ip_input.c Holds ip
destination address for Virtualized. sockaddr_in option processing.
ipflowS static struct ip_flow.c Hash table head for ipflow
Virtualized. ipflowhead structs ipflow_active static int ip_flow.c
Tunable. Enables "fast Invariant. forwarding" flow code.
ipflow_inuse static int ip_flow.c Count of active flow structures.
Virtualized. ipforward_rt static struct route ip_input.c Cached
route for ip forwarding. Virtualized. iforwarding int ip_input.c
Tunable that enabales ip Virtualized. forwarding. ipintrq struct
ifqueue ip_input.c Ip interrupt request queue for Invariant.
incoming packets. Vnet set when packets dequeued. ipport_firstauto
static int ip_pcb.c Bounds on ephemeral ports. Invariant.
ipport_hifirstauto static int ip_pcb.c Bounds on ephemeral ports.
Invariant. ipport_hilastauto static int ip_pcb.c Bounds on
ephemeral ports. Invariant. ipport_lastauto static int ip_pcb.c
Bounds on ephemeral ports. Invariant. ipport_lowfirstauto static
int ip_pcb.c Bounds on ephemeral ports. Invariant.
ipport_lowlastauto static int ip_pcb.c Bounds on ephemeral ports.
Invariant. ipprintfs static int ip_nput.c Flag for debug print.
Invariant. ipq static struct ipq ip_input.c Head of ip reassembly
hash Virtualized. [IPREASS_NHASH] lists. ipqmaxlen static int
ip_input.c Patchable constant that sets Invariant. maximum queue
length for ipintrq. isendredirects static int ip_input.c Tunable
that enable sending Invariant. redirect messages. istaT struct
ipstat ip_input.c Ip statistics counters. Virtualized. k_igmpsrc
static struct ip_mroute.c Prototype sockaddr_in. Invariant.
sockaddr_in last_adjusted_timeout static int ip_rmx.c Time value of
last adjusted Virtualized. timeout. last_encap_src static u_long
ip_mroute.c Cache of last encapsulated Virtualized. source address?
last_encap_vif struct vif* ip_mroute.c Last encapsulated volume tag
Virtualized. (vif). last_zeroed static int radix.c Number of bytes
zeroed last Invariant. time in addmask_key. legal_vif_num int
(*)(int) ip_mroute.c Pointer to function selected by Invariant.
module installation. llinfo_arP struct if_ether.c Head of
llinfo_arp linked list. Virtualized. llinfo_arp_head log_in_vain
static int tcp_input.c Tunables that enable logging of Invariant.
udp_usrreq.c "in vain" connections. loif struct ifnet if_loop.c
Array of ifnet structs fro Invariant. [NLOOP] loopback device. One
per device, therefore invariant. mask_mhead struct radix.c Head of
mask tree. Invariant. radix_node_head* max_keylen static int
radix.c Maximum key length of any Invariant. domain. maxnipq static
int ip_input.c Constant (nmbcluslter/4) that is Invariant? maximum
number of ip Scaled? fragments waiting assembly. Note: should this
be scaled by VNET? mfctable static struct mfc* ip_mroute.c Head of
mfc hash table. Virtualized. [MFCTBLSIZ] mrt_ioctl int (*)(int,
ip_mroute.c Function pointer selected by Invariant. caddr_t struct
module initialization. proc*) mrtdebug static u_int ip_mroute.c
Enables debug log messages. Invariant. mrtstat static struct
ip_mroute.c Multicast routing statistics. Virtualized. mrtstat
mtutab static int [] ip_icmp.c Static table of constants.
Invariant. multicast_decap_if static struct ifnet ip_mroute.c Fake
encapsulator interfaces. Virtualized. [MAXVIFS]
multicast_encap_iphdr static struct ip ip_mroute.c Multicast
encapsulation header. Invariant. nexpire static u_char ip_mroute.c
Count of number of expired Virtualized. [MFCTBLSIZ] entries in hash
table? nipq static int ip_input.c Number of ipfragment chains
Virtualized. awaiting reassembly. normal_chars static char []
radix.c Static table of mask constants. Invariant. nousrreqs static
struct in_proto.c Static structure of null function Invariant.
pr_usrreqs ipx_proto.c pointers. null_sdl.96 static struct
if_ether.c Static null sockaddr_dl Invariant. sockaddr_dl
structure. numvifs static vifi_t ip_mroute.c Number of virtual
interface Virtualized. structures. old_chk_ptr static ip_fw.c
Function pointer holding Invariant. ip_fw_chk_t previous state when
module loads. old_ctl_ptr static ip_fw_ctl_t ip_fw.c Function
pointer holding Invariant. previous state when module loads.
paritytab static unsigned ppp_tty.c Static array of parity
constants. Invariant. [8] pim_assert static int ip_mroute.c Enables
pim assert processing. Virtualized. ppp_compressors static struct
if_ppp.c Static list of known ppp Invariant. compressor [8]
compressors. ppp_softc struct ppp_softc if_ppp.c Array of softc
structures for Invariant. pppdisc [NPPP] ppp driver; one per
device. raw_recvspace static u_long raw_cb.c Patchable constant
that is Invariant. amount of receive space to reserve in socket.
raw_sendspace static u_long raw_cb.c Patchable constant that is
Invariant. amount of send space to reserve in socket. raw_usrreqs
struct protosw raw_usrreq.c Table of function pointers. Invariant.
rawcb_lisT struct raw_cb.c Head of rawcb (raw prototocol
Virtualized. rawcb_list_head control blocks) list. rawclnsdomain
struct domain raw_clns.c Table of function pointers. Invariant.
rip_recvspace static u_long raw_ip.c Tunable, amount of receive
Tunable. Not space to reserve in socket. virtualized. rip_sendspace
static u_long raw_ip.c Tunable, amount of send space Tunable. Not
to reserve in socket. virtualized. rip_usrreqs struct pr_usrreqs
raw_ip.c Table of function pointers. Invariant. ripcb static struct
raw_ip.c Head of raw ipcontrol blocks Virtualized. inpcbhead
ripcbinfo struct inpcbinfo raw_ip.c Pcb info. structure for raw ip.
Virtualized. ripsrc static struct raw_ip.c Static temporary
variable in Invariant. sockaddr_in rip_input m_mkfreelist static
struct radix.c Cache of free radix_mask Invariant. radix_mask*
structures. m_ones static char* radix.c One mask computed from
Invariant. maximum
key length. m_zeros static char* radix.c Zeros mask computed from
Invariant. maximum key length. ro static struct route ip_mroute.c
Temporary variable to hold Invariant. ro route. route_cB struct
route_cb route.c Counts on the number of Virtualized. routing
socket listeners per protocol. route_dst static struct rtsock.c
Null address structure for Invariant. sockaddr route destination.
route_proto static struct rtsock.c Static prototype of structure
Invariant. sockproto used to pass routing info. route_src static
struct rtsock.c Null address structure for Invariant. sockaddr
source. route_usrreqs static struct rtsock.c Table of function
pointers for Invariant. pr_usrreqs entry points. routedomain struct
domain rtsock.c Table of function pointers for Invariant. entry
points. route_alert static struct mbuf* igmp.c Statically
constructed router Invariant. alert option. routesw struct protosw
rtsock.c Table of function pointers for Invariant. entry points.
rsvp_oN int ip_input.c Count of number of open rsvp Virtualized.
control sockets. rsvp_src static struct ip_mroute.c Sockaddr
prototype. Invariant. sockaddr_in rsvpdebug static u_int
ip_mroute.c Enables debug print. Invariant. rt_tableS struct
route.c Head of the routing tables (a Virtualized. radix_node_head*
table per address family.) [AF_AX + 1] rtq_minreallyold static int
in_rmx.c Tunable; minimum time for old Invariant. routes to expire.
rtq_reallyold statinc int in_rmx.c Amount of time before old
Virtualized. routes expire. rtq_timeout static int in_rmx.c
Patchable constant timeout Invariant. value for walking the routing
tree. rtq_toomany static int in_rmx.c Tunable that represents the
Invariant. number of active routes in the tree. rtstaT struct
rtstat route.c Routing statistics structure. Virtualized. rttrash
static int route.c Number of rtentrys not linked Dead code. Not to
the routing table. Never read, virtualized. dead code. sa_zero
struct sockaddr rtsock.c Zero address return in error Invariant.
conditions. sin static struct if_ether.c Sockaddr prototype passed
to Invariant. sockaddr_inarp if_mroute.c rtallocl. sl_softc static
struct if_sl.c Softc structure for slip driver; Invariant. sl_soft
[NSL] one per device. slipdisc static struct if_sl.c Table of
function pointers to Invariant. linesw slip entry points. srctun
static int ip_mroute.c Counter throttling error Invariant. message
to log. subnetsarelocal static int in.c Tunable flag indicating
subnets Virtualized. are local. tbfdebug static u_int ip_mroute.c
Tbf debug level. Invariant. tbftable static struct tbf ip_mroute.c
Token bucket filter structures. Virtualized. [MAXVIFS] tcB struct
inpcbhead tcp_input.c Head structure for tcp pcb Virtualized.
structures. tcbinfO struct inpcbinfo tcp_input.c PCB info structure
for tcp. Virtualized. tcp_backoff int [] tcp_timer.c Table of times
for tcp backff Invariant. tcp_ccgeN tcp_cc tcp_input.c Connection
count (per rfc Virtualized. (u_int32_t) 1644). tcp_delack_enabled
int tcp_input.c Tunable that enables delayed Tunable. Not
acknowledgments. virtualized. tcp_do_rfc1323 static int tcp_subr.c
Tunable enables rcf 1323 Tunable. Not (window scaling and
virtualized. timestamps.) tcp_do_rfc1644 static int tcp_subr.c
Tunable enables rfc 1644. Tunable. Not virtualized. tcp_keepcnt
static int tcp_timer.c Patchable constant for Invariant. maximum
number of probes before a drop. tcp_keepidle int tcp_timer.c
Tunable value for keep alive Tunable. Not idle timer. virtualized.
tcp_keepinit int tcp_timer.c Tunable value for initial Tunable. Not
connect keep alive. virtualized. tcp_maxidle int tcp_timer.c
Product of tcp_keepcnt* Invariant. tcp_keepintvl; recomputed in
show timeout. tcp_maxpersistidle static int tcp_timer.c Patchable
constant that is Invariant. default time before probing.
tcp_mssdflt int tcp_subr.c Tunable default maximum Tunable. Not
segment size. virtualized. tcp_noW u_long tcp_input.c 500 msec.
counter for RFC1323 Virtualized. timestamps. tcp_outflags u_char
tcp_fsm.h Static table of flags in Invariant. [TCP_NSTATES]
tcp_output. tcp_rttdflt static int tcp_subr.c Tunable. Dead code,
value not Invariant. Dead accessed code tcp_sendspace u_long
tcp_usrreq Tunable value for amount of Tunable. Not send space to
reserve on socket. virtualized. tcp_totbackoff static int
tcp_timer.c Sum of tcp_backoff. Invariant. tcp_usrreqs struct
pr_usrreqs tcp_usrreq.c Table of function pointers for Invariant.
tcp user request functions. tcprexmtthresh static int tcp_input.c
Patchable constant; number of Invariant. duplicate acks to trigger
fast retransmit. tcpstaT struct tcpstat tcp_input.c TCP statistics
structure. Virtualized. tun_cdevsw struct cdevsw if_tun.c Table of
function pointers for Invariant. tunnel interface entry points.
tun_devsw_installed static int if_tun.c Flag indiating tun devsw
table Invariant. installed. tunctl static struct if_tun.c Softc
structure for tunnel Invariant. tun_softc interface; one per
device. [NTUN] tundebug static int if_tun.c Flag enables debut
print. Invariant. udb static struct udp_usrreq.c UDP inpcb head
structure. Virtualized. impcbhead udbinfo static struct
udp_usrreq.c UDP inpcb info. structure. Virtualized. impcbinfo
udp_in static struct udp_usrreq.c Prototype sockaddr for Invariant.
sockaddr_in AF_INET. udp_recvspace static u_long udp_usrreq.c
Tunable; amount of receive Tunable. Not space to reserve on socket.
virtualized. udp_sendspace static u_long udp_usrreq.c Tunable;
amount of send space Tunable. Not to reserve on socket.
virtualized. udp_usrreqs struct pr_usrreqs udp_usrreq.c Table of
function pointers for Invariant. entry points. udpcksum static int
udp_usrreq.c Tunable; enables udp Tunable. Not checksumming.
virtualized. udpstat struct udpstat udp_usrreq.c Udp statistics
structure. Virtualized. useloopback static int it_ether.c Tunable;
enables use of Tunable. Not loopback device for localhost.
virtualized. version static int ip_mroute.c Version number of MRT
Invariant. protocol. viftable static struct vif ip_mroute.c Table
of vifs (virtual interface Virtualized. [MAXVIFS] structure).
zeroin_addr struct in_addr in_pcb.c Zero'd internet address.
Invariant.
[0030] Although the present invention and its advantages have been
described in detail, it should be understood that various changes,
substitutions and alterations can be made herein without departing
from the spirit and scope of the invention as defined by the
appended claims. Moreover, the scope of the present application is
not intended to be limited to the particular embodiments of the
process, machine, manufacture, composition of matter, means,
methods and steps described in the specification. As one of
ordinary skill in the art will readily appreciate from the
disclosure of the present invention, processes, machines,
manufacture, compositions of matter, means, methods, or steps,
presently existing or later to be developed that perform
substantially the same function or achieve substantially the same
result as the corresponding embodiments described herein may be
utilized according to the present invention. Accordingly, the
appended claims are intended to include within their scope such
processes, machines, manufacture, compositions of matter, means,
methods, or steps.
* * * * *